Various Williams Classic Blunders

Here’s an array of problems I’ve encountered fixing Defender/Stargate hardware for people.

Rev. C (Sinistar)
No known issues with the MPU. They shipped this with the metal plate you bolt the boards into, they said for safety, and it’s a good thing because I forgot Sinistar’s ROM and input boards had different mounting holes from the other games. This was the first game I tested and it became quite apparent to me that every ROM and input board would need new ribbon cables, which I had just run out of and needed to order more. I’m puzzled that Sinistar came on a Rev. C, cause I thought those were all Rev. D boards. I’ll have to ping a serial expert.

EDIT: I dug out that Ken Graham seminar video and he confirmed all Sinistars came with Rev. D boards. In fact all the boards in this Sinistar set had different serial numbers, so it was cobbled together.

Rev. – (Stargate-only)
Board was previously modified with BS.BA inversion hack (like seen on Rev. B) to work with Robotron/Joust. I’m assuming they stopped when they couldn’t get past the “open coin door” screen. The early Stargate MPUs had an additional daughterboard that plugged into a socket at chip location 6D with 2 jumper wires spanning from it. This board is cited in the Stargate Drawing Set as IC1, if you look around by the 6514 CMOS part IC1 will be enclosed in a dashed line pattern. All parts in the schematic inside the dashed line pattern are physically located on the daughterboard itself. IC1 is necessary on Rev. – boards as they were printed without all the necessary parts for the CS NOT and WE functions of the CMOS on the MPU itself; IC1 reimplements these missing components. Without IC1 installed the CMOS will not initialize properly, leaving you stranded on the “open coin door” screen.

I attempted to populate and jumper all these parts myself and was unable to do so with what I had available. I had a couple Rev. – boards in my box of stuff with IC1 boards, so I just donated one of those to the cause to make this one work.

Rev. C (battery leak)
Another Rev. C board with battery leakage. I’m uncertain as to why seemingly every one of these I’ve come across have had this problem. There are two separate effects of old batteries: 1) the alkaline will leak out and get into the traces and 2) they can outgas, creating a chemical reaction with nearby solder or components. This board had alkaline in the thick ground trace and visibly the solder looked like concrete on the row 7 chips (where clock generation is created). If the solder has been compromised, you may find it next to impossible to get new solder to stick.

So my first step was taking the board to the sink and dumping some Zep Acidic Toilet Bowl cleaner on the affected areas and using a toothbrush to circulate the Zep around. The solder mask on the areas where alkaline is embedded in the traces will generally just flake off with the toothbrush or if you really want to get in, a flathead screwdriver will definitely make it come off. There were other chips that looked browned with rust and other gunk in other areas I cleaned up too. Immediately after I rinsed the Zep off and took to giving it a Simple Green wash in the bathtub next. Drying it out with a hair dryer after, and some additional sanding down to the bare copper, the clean up is complete. You don’t want to leave any significant amount of alkaline in the traces because it will spread and eventually the traces will disintegrate.

Next I proceeded to desolder out and replace the 7J (74LS107) and 7K (7404) chips. I settled on replacing 7F and 7G (7474s) as well for added measure. The board still had no reset, pin 37 of the CPU was showing Low on the logic probe. After more research I came across transistor Q12 (2N4403), which I’ve never seen fail in the past, and wound up summoning a replacement off a Defender MPU I knew worked. I put the battery board Q12 in the Defender and powered that up first, pin 37 of the CPU was Low; it appeared I found the bad part. I installed the donor Q12 in and now I have life out of the battery leak board.

Not done yet however, the rug pattern had an overlapping cascading effect to it. That stuck out to me as multiplexer failure. The boards have 4 of these 74153 parts on them, and the traditional test of the outputs at pins 7 and 9 with the logic probe was of no use here because they were pulsing just fine. I was able to piggyback new 74153s on chips that had failed in the past, but I think that was when the outputs were floating. In this instance you keep piggybacking multiplexers until you either make it look worse on the rug pattern screen, or if it makes any kind of jittering stop. As such I was able to find the bad parts and replaced them.

I was left with an array of assorted RAM errors afterwards. I kept swapping chips around until I identified a few bad ones (4 in all, if I recall). I narrowed it down to a consistent 1-3-8 error, popping good chips in didn’t change anything. I lifted the plastic housing off the socket for RAM 38 and it was revealed underneath that the pins were all green. I replaced the socket, now I’m back to 1-3-1. It turned out when I reinstalled the RAMs I had the +5 pin of the chip hanging out of the socket. (RAM 35 if I recall) I was able to deduce this logic probing the parallel buffer 74374 at 4H and found the D36 line wasn’t giving a clean pulse, leading me to find which RAM was causing the problem. Upon sorting that out now the board works 100%. What an endeavor.

Rev. B (Robotron possibly)
This board had an ugly rug pattern like the battery leak board. In addition to the rug pattern producing an overlapping cascading effect, it also had a jittering effect up and down too. I had a bunch of 74153 chips, I just piggybacked new ones on top of the 4 on the board. The jittering stopped. I was able to narrow down which chip was bad after I started pulling the piggyback chips off and got the jitter to come back. This one also had good pulses on the outputs at pins 7 and 9. When I started fixing these years ago my early repair boards had bad multiplexers, so I wound up stocking up on tons of them. Then suddenly it wasn’t a common issue anymore for years.

Failures of other boards:

Joust ROM Board
Immediately recognized the ribbon cable was bad on this board, as it required a fair amount of bending around and pressing on it to make the rug pattern come up. While trying to load the game however, a glance at the error display on the board showed a bunch of random numbers popping up, and ultimately ending with a U when the game should’ve been on the high score table. The game would then reset. Both symptoms are signs of PIA failure, as that controls the output to the error display and the game crash resetting is probably a failure of Count 240 functioning correctly. I have also seen the 4049 chip cause the U error, but the game would lock up to a black screen in this instance. I socketed and replaced the PIA, now it behaves like it should.

Robotron ROM Board
Another bad ribbon cable, you can wiggle it around and the game will boot up just fine.

Sinistar ROM Board
Yes, another bad ribbon cable. Like Robotron, it booted up just fine. But the Auto Up/Manual Down switch was not registering at all. One of the green square coil packs had a pin that came unsoldered. There was no signal at all at the header pin. It wound up being a bad 4049 chip, which I had guessed was the root cause but I wound up taking the thorough long way approach to going over every other avenue.

Stargate Sound Board
Plugged this sound board in and was greeted to an array of beeping sounds. Interestingly this had a speech header on it (like for the talky pinball games/Sinistar) which means it was probably from a Defender. Someone attempted to reflow the solder to the speech header and did it very haphazardly. I wound up desoldering it off as it’s unneeded for Stargate anyway and now it goes womp womp again.

Joust Sound Board
I had no audio output at all from this, but the amp was functional cause I could make it buzz and whir depending on various parts I put my fingers on. I was led to believe that PIA failure would only result in wrong sound playback, but today I was wrong. I probed the address and data lines at the ROM and D0 came up a little wrong. I traced this back to the 1408 chip where I found some floating pins with the probe and TTL tested it against a working sound board and found that chip was in fact bad. Replacing it didn’t restore sound though. It wound up being the 6821 PIA was bad. Now you can hear thumping feet sounds with the test switch. I didn’t suspect PIA initially because the test switch (I was told) is supposed to bypass the PIA. I’ve learned the complex way that’s not necessarily true, unless this PIA was bad in the way it was just conflicting with the buses that much.

Sinistar Sound Board
Everything from a logic standpoint worked correctly, but it was very quiet. The Balance pot on the speech board seemed dirty because I turned it up a little bit and the volume shot way up. Sprayed it down with contact cleaner, should be fine now. I’ll have to notate this on the Sound Board page, but a replacement Balance pot can be obtained here.

Joust ROM Board
This was moderately confusing at first because the MPU I paired it with had an unusual screen. Typically the MPU will display a bar pattern if the ROM board is unplugged from power or the ribbon cable is bad, but this was just like a dark grey screen with random yellow and blue dots. I thought initially it was just a bad ribbon cable (which it probably was) but replacing it resulted in no change. I decided to take a break for the night and come back to it the next day. Here I tried running it with the ROM board unplugged and to my surprise, the pattern on the screen stayed the same. That’s just how that MPU looked. So I plugged the other MPU in and got the more familiar green and blue bar screen. Connected the ROM board, same exact pattern. It has a new ribbon cable, I don’t understand this. After deducing the ribbon cable was working just fine, I started trying to hunt around other repair logs to find a similar symptom. I came up empty. Entirely on a hunch I socketed and replaced the 74154 decoder/demultiplexer chip and it booted right up. So if you have a ROM board that looks like it’s just unplugged, that decoder/demultiplexer chip might be bad.

Other Joust ROM Board
Typical ribbon cable failure, except this time the replacement was one of those PC IDE cable module gadgets. Which was helpful, there’s 4 Joust ROM boards in all here at once, and I haven’t gotten around to labeling them yet, so this one stuck out like a sore thumb. I found no other issues with it, it works.


The battery leak legend of “Joust Foley”. I apparently got impatient and immediately started to trying to clean up the mess, so there’s no real “before” pictures. I imagine the battery holder looked like a disaster. As this came out of a cocktail machine and they mount the boards a little differently, the damage reached the opposite side on the power header end. This was a dead board if I recall. Desoldering the 74LS107 chip out wound up taking all the top solder pads with it. This and the 7404 chip near the battery holder are typically the ones that get eaten up in the event of outgassing or alkaline leakage. This particular board had more outgassing than anything, I didn’t really need to sand anything else besides the chip traces and the battery holder pads. I wound up just donating a replacement board to the cause, and this one I traded later.

This was a different board, but equally crusty results.

Here was a Defender power supply where you could see the header solder cracked (or haloed) in multiple spots. I’m guessing it got some Bob Roberts treatment previously as the 18,000 uf capacitor was replaced with a radial with jumper wires. I don’t remember whose this was.

I don’t like the original bridge rectifiers with tabs mounted low against the board. Never attempt to desolder a bridge rectifier, as those 4 diodes inside will just absorb the soldering iron heat. The heat will collect in the solder pads and they will probably more than likely eventually fall off in the midst of your removal efforts. Instead cut the diode package from the legs, then you can plunk the legs out and clean the solder with much greater ease.

The original header pins Williams used are inferior as they’re round and skinny, and will thus be inadequate with new trifurcon pins in the plugs because they won’t be able to grab the pins. I replace them with these black based headers. Since they’re square and way more stout, you’ll find removing plugs to be way easier as the pins won’t bend easily side to side like the old round ones.

That’s much better. Hopefully it still works well for whoever it belongs to.

Here was a Joust that had an issue with the vertical Decoder 6. Maybe this was “Joust Foley” too. If it appears the image is repeating vertically then you might have a broken socket. Popping the housing off I found a broken pin inside.

I don’t seem to have the “after” picture but I can assure you it worked after replacing that socket.

This was a 6821 PIA on I’m estimating was a Robotron ROM board. If your game has no sound output at all, a scrambled error display, the game has some graphical anomalies, or it crash resets after UNIT OK/ALL SYSTEMS GO, suspect a bad ROM PIA. If you desolder a chip out and see scorch marks on the board like seen here, chances are it’s toast.

I may add some more nuggets to this later as I find them, my picture archive is pretty deep and that might take awhile.